What is the difference between channelization codes and scrambling codes in UMTS CDMA?

Channelization codes (OVSF) separate users within the same cell; scrambling codes separate cells (and sectors) from each other. The signal is first spread with OVSF, then scrambled.

* Spread-then-scramble: OVSF separates users, scrambling separates cells. *

Two-stage process:

User data → × Channelization code (OVSF) → × Scrambling code → Transmitted signal

Channelization codes (OVSF):

• Separate users within one cell/sector
• Perfectly orthogonal — zero cross-correlation → no intra-cell interference
• Generated from the OVSF tree
• Each user gets a unique OVSF code from the same tree

Scrambling codes:

• Separate cells and sectors from each other
• Pseudo-orthogonal (pseudo-random, not perfectly orthogonal) → some residual inter-cell interference
• Each base station sector gets a unique scrambling code
• All OVSF-spread signals from one sector are combined, then multiplied by the sector's scrambling code

Why two layers?

• Within a cell, you want perfect separation → orthogonal OVSF codes
• Between cells, perfect orthogonality is impossible (no synchronization between base stations), so pseudo-random scrambling provides good enough separation
• The remaining inter-cell interference is managed by power control

Result: A user's signal is orthogonal to other users in the same sector, and pseudo-orthogonal to signals from other sectors and base stations.

Go deeper:

• UMTS WCDMA — scrambling codes, uplink vs downlink (ShareTechnote) — the spread-then-scramble chain in detail, including why downlink uses long codes and uplink short codes, with constellation views of the combined channels.